This renewal application requests 2.5 years of support to undertake the late stages of a prevalence and incidence study of Alzheimer's disease (AD) in relation to apolipoprotein E (APOE) polymorphisms and a variety of health history and environmental variables. The first 5 years of this work is nearing completion, but our present 5-year award could not provide funds for the final stages of work, proposed here. These include follow-up examination of newly identified incident cases 18 months after detection, and analyses and write up of the data. We have also added a new specific aim that will provide more objective, algorithmic diagnoses of dementia from standardized psychometric tests. This method will be applied to data from the Prevalence Wave, as well as incidence cases, to help avoid drift in diagnoses. It will also improve the comparability of our findings with those from other studies. The sample is the entire elderly population of Cache County, Utah (5,713 eligible, 5,092 participants, 90% response rate). Among respondents, 731 were age 85 or older and 254 age 90 or older. An estimated 4,400 subjects were not demented when assessed in 1995-96 and will participate in Incidence Wave screening (currently in progress). Participation rates at all stages are unusually high (85% completion rate for dementia evaluations among Prevalence Wave participants). The emphasis is on rates among the oldest old, and on identification and characterization of factors that may delay onset of AD, and thus reduce its prevalence. Nested prevalent case-control findings indicate possible protective effects of non-steroidal anti-inflammatory drugs, histamine H2 blockers, and dietary anti-oxidants. These exposures will receive emphasis in the incidence phase of the study as a likely preface to randomized primary prevention trials.